Drink and Thrive?
German Economics of Binge-Drinking

CSSR Presentation

Structure

• Introduction
• Data
• Descriptive Statistics
• Empirical Strategy
• Results

Trend Sport of the 21st Century?

Background

Binge Drinking – A Problem?

German crime statistics: violent crime increasingly committed under the influence of alcohol

Adolescents especially affected

Research Questions

Similar to Marcus and Siedler, 2015:

What explains regional differences? Correlation with economic situation?

Night sales ban on alcohol: any effect on binge drinking in Baden-Wuerttemberg?

Data

Variables	Reasoning
F10.0 cases per 1000	Number of people hospitalized with acute alcohol intoxication
K70 cases per 1000	Number of people hospitalized diagnosed with alcohol-related liver disease
GDP per capita	Value and change between years as indicator of economic development
(Youth) unemployment rate	Indicator of economic situation for age groups
Population density	Indicator for urbanization
Treatment dummies (local & temporal)	Capture the policy measure effect

Data Visuals - F10.0 per Age Group

Data Visuals - K70 per Age Group

Data Visuals - F10.0 Cases in Germany

Data Visuals - F10.0 Increase Over Time

Data Visuals - F10.0 Increase - 15-19 y/o

Empiricial Strategy

• Dependent variable: Alcohol-related hospitalization (per 1000)

Question 1: Differences between states

Method: Multiple regression

• Variation: Levels and differences
• Robustness check: Different years and long vs. short-term alcohol diagnosis

Question 2: Effect of night-sale-ban

Method: Difference-in-Difference

• Variation: Control variables and number of periods
• Robustness check: different control group compositions

Results Question 1

Result Night Sale Ban

F10.0 Diagnoses for 15-19 year olds (control: 13 non-city states)

Model 5: multi-period panel differences-in-differences

	Dependent variable:
	F10.0 cases per 1000 PP
BW night sale ban	-0.06***
	(0.01)
Youth unemployment rate	0.02***
	(0.002)
GDP per capita	0.01***
	(0.003)
Observations	195
Adjusted R2	0.42
Note:	p<0.1; p<0.05; p<0.01

Conclusion

Marcus and Sielder (2014) find about 7% decrease, we find about 20% decrease

Conclusion 2

Problems of our research design:

• Availability of data
  • without paying only annual and state-level data available
  • 14 years x 16 states = 224 obs.
  • access to monthly and county level data would allow
  • 168 months x 402 counties = 67536 obs.
• limited amount of control variables
  • few meaningful state level variables available
  • but: less of a problem in Diff-in-Diff method

Appendix 1

Regression results for Model 1 for 2000, 2007 and 2014

	Dependent variable:
	F10.0 Diagnoses per 1000 capita
	2000	2007	2014
GDP per capita	-0.02**	-0.01	-0.01
	(0.01)	(0.01)	(0.01)
Unemployment rate	-0.04***	-0.02	-0.03
	(0.01)	(0.03)	(0.04)
Beer tax	0.0001*	-0.0002	-0.0002*
	(0.0001)	(0.0001)	(0.0001)
Population density	1.63***	1.95***	2.23***
	(0.31)	(0.58)	(0.59)
Observations	16	16	16
Adjusted R2	0.36	0.40	0.48
Residual Std. Error (df = 12)	0.15	0.28	0.28
Note:	p<0.1; p<0.05; p<0.01

Appendix 2

Regression results for Model 2 with first differenced data

	Dependent variable:
	Change in F10.0	Change in F10.2	Change in K70
GDP Change	0.02***	-0.01	0.002
	(0.01)	(0.01)	(0.002)
Unemployment Change	-0.01	-0.03**	0.002
	(0.01)	(0.01)	(0.003)
Observations	224	224	224
Adjusted R2	0.08	0.02	-0.005
Residual Std. Error (df = 222)	0.11	0.18	0.04
Note:	p<0.1; p<0.05; p<0.01

Appendix 3

Model 3: simple differences-in-differences

	Dependent variable:
	F10.0 cases per 1000 PP
	All States	All but City States
BW night sale ban	0.12	0.04
	(0.12)	(0.11)
Post-treatment dummy	0.43***	0.48***
	(0.05)	(0.05)
Interaction	-0.11	-0.16
	(0.21)	(0.18)
(Intercept)	0.96***	1.04***
	(0.03)	(0.03)
Observations	255	210
Adjusted R2	0.22	0.31
Residual Std. Error	0.37 (df = 251)	0.32 (df = 206)
Note:	p<0.1; p<0.05; p<0.01

Appendix 4

Model 4: simple diff-in-diff with controls

	Dependent variable:
	F10.0 cases per 1000
BW night sale ban	-0.16
	(0.11)
Post-sales ban dummy	0.32***
	(0.05)
GDP per capita	-0.02***
	(0.003)
Youth unemployment rate	-0.06***
	(0.01)
Interaction dummy	0.04
	(0.18)
(Intercept)	2.30***
	(0.13)
Observations	240
Adjusted R2	0.45
Residual Std. Error	0.32 (df = 234)
Note:	p<0.1; p<0.05; p<0.01

Appendix 5

Model 5: multi-period panel differences-in-differences

	Dependent variable:
	F10.0 cases per 1000 PP
	All States	15-19y	15-19y + control: non-city
BW night sale ban	-0.06	-0.03*	-0.06***
	(0.09)	(0.02)	(0.01)
Youth unemployment rate	-0.02**	0.02***	0.02***
	(0.01)	(0.002)	(0.002)
GDP per capita	-0.01	0.003	0.01***
	(0.02)	(0.003)	(0.003)
Observations	240	240	195
Adjusted R2	0.02	0.29	0.42
Note:	p<0.1; p<0.05; p<0.01

Appendix 6